Accelerating frequency-domain diffuse optical tomographic image reconstruction using graphics processing units

A simple algorithm for diffuse optical tomography without Jacobian inversion

A computationally simpler algorithm to reconstruct the optical property distribution of turbid media using diffuse optical tomographic principles is presented. The proposed algorithm eliminates the requirement of large Jacobian matrix inversion which otherwise is essential for tomographic imaging. The most significant Jacobians are identified based on proper thresholding of the measurement and the intersection of these Jacobians gives the approximate spatial location of the inhomogeneity. The algorithm is tested and optimized using simulations and further validated using tissue-mimicking phantom-based experiments andin-vivosmall-animal experiments.

Toward real-time diffuse optical tomography: accelerating light propagation modeling employing parallel computing on GPU and CPU

Parameter recovery in diffuse optical tomography is a computationally expensive algorithm, especially when used for large and complex volumes, as in the case of human brain functional imaging. The modeling of light propagation, also known as the forward problem, is the computational bottleneck of the recovery algorithm, whereby the lack of a real-time solution is impeding practical and clinical applications. The objective of this work is the acceleration of the forward model, within a diffusion approximation-based finite-element modeling framework, employing parallelization to expedite the calculation of light propagation in realistic adult head models. The proposed methodology is applicable for modeling both continuous wave and frequency-domain systems with the results demonstrating a 10-fold speed increase when GPU architectures are available, while maintaining high accuracy. It is shown that, for a very high-resolution finite-element model of the adult human head with ∼600,000 nodes, consisting of heterogeneous layers, light propagation can be calculated at ∼0.25  s/excitation source.

GPU-based acceleration and mesh optimization of finite-element-method-based quantitative photoacoustic tomography: a step towards clinical applications

Finite element method (FEM)-based time-domain quantitative photoacoustic tomography (TD-qPAT) is a powerful approach, as it provides highly accurate quantitative imaging capability by recovering absolute tissue absorption coefficients for functional imaging. However, this approach is extremely computationally demanding, and requires days for the reconstruction of one set of images, making it impractical to be used in clinical applications, where a large amount of data needs to be processed in a limited time scale. To address this challenge, here we present a graphic processing unit (GPU)-based parallelization method to accelerate the image reconstruction using FEM-based TD-qPAT. In addition, to further optimize FEM-based TD-qPAT reconstruction, an adaptive meshing technique, along with mesh density optimization, is adopted. Phantom experimental data are used in our study to evaluate the GPU-based TD-qPAT algorithm, as well as the adaptive meshing technique. The results show that our new approach can considerably reduce the computation time by at least 136-fold over the current central processing unit (CPU)-based algorithm. The quality of image reconstruction is also improved significantly when adaptive meshing and mesh density optimization are applied.

Parallel Solver for Diffuse Optical Tomography on Realistic Head Models With Scattering and Clear Regions

Diffuse optical tomography is an imaging technique, based on evaluation of how light propagates within the human head to obtain the functional information about the brain. Precision in reconstructing such an optical properties map is highly affected by the accuracy of the light propagation model implemented, which needs to take into account the presence of clear and scattering tissues. We present a numerical solver based on the radiosity-diffusion model, integrating the anatomical information provided by a structural MRI. The solver is designed to run on parallel heterogeneous platforms based on multiple GPUs and CPUs. We demonstrate how the solver provides a 7 times speed-up over an isotropic-scattered parallel Monte Carlo engine based on a radiative transport equation for a domain composed of 2 million voxels, along with a significant improvement in accuracy. The speed-up greatly increases for larger domains, allowing us to compute the light distribution of a full human head ( ≈ 3 million voxels) in 116 s for the platform used.

Mesh Optimization for Monte Carlo-Based Optical Tomography

Mesh-based Monte Carlo techniques for optical imaging allow for accurate modeling of light propagation in complex biological tissues. Recently, they have been developed within an efficient computational framework to be used as a forward model in optical tomography. However, commonly employed adaptive mesh discretization techniques have not yet been implemented for Monte Carlo based tomography. Herein, we propose a methodology to optimize the mesh discretization and analytically rescale the associated Jacobian based on the characteristics of the forward model. We demonstrate that this method maintains the accuracy of the forward model even in the case of temporal data sets while allowing for significant coarsening or refinement of the mesh.

Towards real-time detection of seizures in awake rats with GPU-accelerated diffuse optical tomography

BACKGROUND

Advancement in clinically relevant studies like seizure interruption using functional neuro imaging tools has shown that specific changes in hemodynamics precede and accompany seizure onset and propagation. However, preclinical seizure experiments need to be conducted in awake animals with images reconstructed and displayed in real-time.

METHODS

This article describes an approach that can be utilized to tackle these challenges. A subject specific head interface and restraining method was designed to allow for DOT to imaging of hemodynamic changes in unanesthetized rats during evoked acute seizures. Using CUDA programming model, the finite-element based nonlinear iterative algorithm for image reconstruction was parallelized.

RESULTS

Early hemodynamic changes were monitored in real time and observed tens of seconds prior to seizure onset. Utilizing the massive parallelization offered by graphic processing units (GPU), DOT was extended to online image reconstruction within 1s.

COMPARISON WITH EXISTING METHODS

Pre-seizure state related hemodynamic changes were detected in awake rats. 3D monitoring of hemodynamic changes was performed in real time with our parallelized image reconstruction procedure.

CONCLUSION

Diffuse optical tomography (DOT) is a promising neuroimaging tool for the investigation of seizures in awake animals.

High-Speed GPU-Based Fully Three-Dimensional Diffuse Optical Tomographic System

We have developed a graphics processor unit (GPU-) based high-speed fully 3D system for diffuse optical tomography (DOT). The reduction in execution time of 3D DOT algorithm, a severely ill-posed problem, is made possible through the use of (1) an algorithmic improvement that uses Broyden approach for updating the Jacobian matrix and thereby updating the parameter matrix and (2) the multinode multithreaded GPU and CUDA (Compute Unified Device Architecture) software architecture. Two different GPU implementations of DOT programs are developed in this study: (1) conventional C language program augmented by GPU CUDA and CULA routines (C GPU), (2) MATLAB program supported by MATLAB parallel computing toolkit for GPU (MATLAB GPU). The computation time of the algorithm on host CPU and the GPU system is presented for C and Matlab implementations. The forward computation uses finite element method (FEM) and the problem domain is discretized into 14610, 30823, and 66514 tetrahedral elements. The reconstruction time, so achieved for one iteration of the DOT reconstruction for 14610 elements, is 0.52 seconds for a C based GPU program for 2-plane measurements. The corresponding MATLAB based GPU program took 0.86 seconds. The maximum number of reconstructed frames so achieved is 2 frames per second.

Model-resolution-based basis pursuit deconvolution improves diffuse optical tomographic imaging

The image reconstruction problem encountered in diffuse optical tomographic imaging is ill-posed in nature, necessitating the usage of regularization to result in stable solutions. This regularization also results in loss of resolution in the reconstructed images. A frame work, that is attributed by model-resolution, to improve the reconstructed image characteristics using the basis pursuit deconvolution method is proposed here. The proposed method performs this deconvolution as an additional step in the image reconstruction scheme. It is shown, both in numerical and experimental gelatin phantom cases, that the proposed method yields better recovery of the target shapes compared to traditional method, without the loss of quantitativeness of the results.

Three-dimensional photoacoustic tomography based on graphics-processing-unit-accelerated finite element method

Compared with commonly used analytical reconstruction methods, the frequency-domain finite element method (FEM) based approach has proven to be an accurate and flexible algorithm for photoacoustic tomography. However, the FEM-based algorithm is computationally demanding, especially for three-dimensional cases. To enhance the algorithm's efficiency, in this work a parallel computational strategy is implemented in the framework of the FEM-based reconstruction algorithm using a graphic-processing-unit parallel frame named the "compute unified device architecture." A series of simulation experiments is carried out to test the accuracy and accelerating effect of the improved method. The results obtained indicate that the parallel calculation does not change the accuracy of the reconstruction algorithm, while its computational cost is significantly reduced by a factor of 38.9 with a GTX 580 graphics card using the improved method.

Medical image processing on the GPU - past, present and future

Graphics processing units (GPUs) are used today in a wide range of applications, mainly because they can dramatically accelerate parallel computing, are affordable and energy efficient. In the field of medical imaging, GPUs are in some cases crucial for enabling practical use of computationally demanding algorithms. This review presents the past and present work on GPU accelerated medical image processing, and is meant to serve as an overview and introduction to existing GPU implementations. The review covers GPU acceleration of basic image processing operations (filtering, interpolation, histogram estimation and distance transforms), the most commonly used algorithms in medical imaging (image registration, image segmentation and image denoising) and algorithms that are specific to individual modalities (CT, PET, SPECT, MRI, fMRI, DTI, ultrasound, optical imaging and microscopy). The review ends by highlighting some future possibilities and challenges.

Toward real-time availability of 3D temperature maps created with temporally constrained reconstruction

PURPOSE

To extend the previously developed temporally constrained reconstruction (TCR) algorithm to allow for real-time availability of three-dimensional (3D) temperature maps capable of monitoring MR-guided high intensity focused ultrasound applications.

METHODS

A real-time TCR (RT-TCR) algorithm is developed that only uses current and previously acquired undersampled k-space data from a 3D segmented EPI pulse sequence, with the image reconstruction done in a graphics processing unit implementation to overcome computation burden. Simulated and experimental data sets of HIFU heating are used to evaluate the performance of the RT-TCR algorithm.

RESULTS

The simulation studies demonstrate that the RT-TCR algorithm has subsecond reconstruction time and can accurately measure HIFU-induced temperature rises of 20°C in 15 s for 3D volumes of 16 slices (RMSE = 0.1°C), 24 slices (RMSE = 0.2°C), and 32 slices (RMSE = 0.3°C). Experimental results in ex vivo porcine muscle demonstrate that the RT-TCR approach can reconstruct temperature maps with 192 × 162 × 66 mm 3D volume coverage, 1.5 × 1.5 × 3.0 mm resolution, and 1.2-s scan time with an accuracy of ±0.5°C.

CONCLUSION

The RT-TCR algorithm offers an approach to obtaining large coverage 3D temperature maps in real-time for monitoring MR-guided high intensity focused ultrasound treatments.

Effective contrast recovery in rapid dynamic near-infrared diffuse optical tomography using ℓ(1)-norm-based linear image reconstruction method

Traditional image reconstruction methods in rapid dynamic diffuse optical tomography employ ℓ(2)-norm-based regularization, which is known to remove the high-frequency components in the reconstructed images and make them appear smooth. The contrast recovery in these type of methods is typically dependent on the iterative nature of method employed, where the nonlinear iterative technique is known to perform better in comparison to linear techniques (noniterative) with a caveat that nonlinear techniques are computationally complex. Assuming that there is a linear dependency of solution between successive frames resulted in a linear inverse problem. This new framework with the combination of ℓ(1)-norm-based regularization can provide better robustness to noise and provide better contrast recovery compared to conventional ℓ(2)-based techniques. Moreover, it is shown that the proposed ℓ(1)-based technique is computationally efficient compared to its counterpart (ℓ(2)-based one). The proposed framework requires a reasonably close estimate of the actual solution for the initial frame, and any suboptimal estimate leads to erroneous reconstruction results for the subsequent frames.

High-performance image reconstruction in fluorescence tomography on desktop computers and graphics hardware

Image reconstruction in fluorescence optical tomography is a three-dimensional nonlinear ill-posed problem governed by a system of partial differential equations. In this paper we demonstrate that a combination of state of the art numerical algorithms and a careful hardware optimized implementation allows to solve this large-scale inverse problem in a few seconds on standard desktop PCs with modern graphics hardware. In particular, we present methods to solve not only the forward but also the non-linear inverse problem by massively parallel programming on graphics processors. A comparison of optimized CPU and GPU implementations shows that the reconstruction can be accelerated by factors of about 15 through the use of the graphics hardware without compromising the accuracy in the reconstructed images.